List your company

Instruction Following on ALFWorld Unseen (val)

79.19Success Rate (SR)TCOD (F2B)
-0.83819.938540.71561.4915Apr 27, 2026

Evaluation results

MethodLinks
TCOD (F2B)Student Model=Qwen2.5-3B2026.0479.1912.47
TCOD (B2F)Student Model=Qwen2.5-7B2026.0477.6113.16
Qwen2.5-7B-RL (Teacher)Model Type=Teacher2026.0476.8713.06
TCOD (F2B)Student Model=Qwen2.5-7B2026.0476.1213.22
Vanilla OPDStudent Model=Qwen2.5-7B2026.0472.1413.37
TCOD (B2F)Student Model=Qwen2.5-3B2026.0470.914.56
Vanilla OPDStudent Model=Qwen2.5-3B2026.0460.4516.21
SFTStudent Model=Qwen2.5-7B2026.0448.7320.11
SFTStudent Model=Qwen2.5-3B2026.0425.3724.16
Zero-ShotStudent Model=Qwen2.5-7B2026.048.9628.46
Zero-ShotStudent Model=Qwen2.5-3B2026.042.2429.63